NICE RAPID REVIEW. Workplace policies and interventions for Smoking Cessation. Final Draft. September 2006

NICE RAPID REVIEW Workplace policies and interventions for Smoking Cessation Final Draft September 2006 Kirsten Bell, PhD Lucy McCullough, BSc. Karen DeVries, MSc. Lorraine Greaves, PhD Natasha Jategaonkar, MSc. 1

Table of Contents 1. Executive Summary...3 EVIDENCE STATEMENTS...5 2. Background...14 2.1 Current Smoking Restrictions in England...14 2.2 Smoking-Related Inequalities...16 2.3 Smoke-free Legislation in England...16 3.1 Literature Search...18 3.2 Selection of Studies for Inclusion...18 3.2.1 Policies and Interventions...18 3.3 Outcomes...18 39 studies met inclusion criteria...19 3.4 Quality Appraisal...19 3.5 Synthesis...21 4. Summary of Findings...22 4.1 WORKPLACE INTERVENTIONS...24 4.1.1 What is the effectiveness of smoking cessation support and services to smokers in the context of a smoke-free workplace?...24 4.1.2 Which interventions in the workplace work best?...26 4.1.3 To what extent does the type of workplace and/or nature of employment influence smoking cessation outcomes?...31 4.1.4 How does the effectiveness of workplace smoking interventions vary with factors such as age, sex, gender, class or ethnicity?...32 4.1.5 What are the adverse or unintended outcomes of the intervention?...32 4.1.6 How acceptable is the intervention to those affected by it?...33 4.2 WORKPLACE POLICY...33 4.2.1 What is the effectiveness of the smoke-free workplace policy in:...34 4.2.2 How does the way that workplace policies are implemented influence effectiveness?...38 4.2.3 What is the impact of smoke-free workplaces on uptake of smoking cessation resources? For example, NHS Stop Smoking Services, telephone quitlines, etc....38 4.2.4 What steps could be taken prior to the introduction of smoke-free regulations to maximise the impact? E.g., public information campaigns, increased resources for smoking cessation support...42 4.2.5 How acceptable is the policy to people affected by it?...43 4.2.6 What factors affect compliance?...47 4.2.7 What are the adverse or unintended outcomes of the policy?...48 5. Overview and Discussion...53 6. EVIDENCE TABLE...58 7. APPENDIX A Excluded Studies...87 8. APPENDIX B Background Papers...89 9. APPENDIX C...91 10. APPENDIX D...113 2

1. Executive Summary This review contains assessments of the available data on the extent to which workplace policies stimulate, support and utilise smoking cessation, with a view to determining the likely effects of the introduction of national smoke-free legislation in England in 2007. The available data has been assessed to answer 13 preset questions examining in detail the effectiveness of both workplace policies and workplace interventions in facilitating smoking cessation. Method: A comprehensive literature search was conducted. A total of 13,023 titles and abstracts were screened and full paper copies of 57 studies were assessed for inclusion as evidence, with 39 studies identified as direct evidence. Results: Cessation programmes aimed at the individual when combined with an institutional approach that provides environmental support for stopping smoking are effective in facilitating smoking cessation and smoking restrictions positively impact the uptake of smoking cessation resources. The most effective workplace interventions are those with proven effectiveness in other settings. While financial incentives do not appear to increase the quit rates of these interventions, they can improve recruitment rates, thereby leading to higher absolute numbers of quitters in the long-term. Social or buddy support seems to have a limited effect on workplace interventions, although intensive interventions are more effective than minimal interventions in facilitating smoking cessation. Workplace health assessments with feedback may also have a role to play, although evidence on their impact on smoking cessation amongst employees is currently inconclusive. Alternative methods for smoking cessation that may have a role to play in the workplace are Allen Carr seminars, online smoking cessation support, and integrated smoking cessation and occupational health and safety programmes. However, evidence on the effectiveness of these intervention types is presently weak and further research is needed to determine their effectiveness There are few available studies on how the nature of employment and socio-demographic factors influence cessation outcomes in workplace interventions. Although there is some evidence that managers are more likely to successfully quit smoking, this may be partly explained by baseline differences between managers and other job categories. There are clearly gender differences in people s approaches to smoking cessation although quitting success is reported to be comparable. Given that smoke-free legislation operates simultaneously within a comprehensive tobacco policy context, it is extremely difficult to disentangle the specific impacts of each policy on populations or sub-populations. There is very strong evidence that workplace policies decrease the incidence of smoking in the workplace, thereby reducing the exposure of employees to ETS at work. However, the evidence that overall daily cigarette consumption and smoking prevalence decrease as a result of workplace smoking bans is inconclusive although early reports from countries that have implemented smoke-free legislation are promising. It is possible that workplace smoking bans may have a reduced impact on people with a lower education and/or women of low socio-economic status. Attitudes towards smoking bans appear to vary based on the setting of the policy attitudes to workplace and restaurant bans are more favourable than attitudes towards bans in bars and pubs although attitudes become more favourable following the implementation of such bans. Indeed, evidence from countries that have implemented smoke-free legislation indicates that support for the legislation increases significantly following its implementation. Based on indicators such as education and occupation, people from lower SES groups and bar and restaurant managers appear to be the most opposed to bans in pubs/bars. According to monitoring data from countries that have implemented smoke-free legislation, compliance with the legislation is extremely high. However, varying degrees of compliance in bars appear to be associated with their socioeconomic and demographic characteristics. Although the benefits of workplace smoking bans significantly outweigh their side effects, several adverse and unintended outcomes may potentially accompany the introduction of 3

smoking bans, such as unhealthy smoking behaviours although the extent to which these smoking patterns undermine the net health benefits of smoking bans has not been established. Workplace smoking bans may also increase tensions between smokers and non-smokers, increase perceived exposure to ETS because of intensified contact with smoking at entrances and exits to buildings and may also lead to unsafe smoking practices. 4

EVIDENCE STATEMENTS Note to Reader: There are two types of evidence drawn on in this review: 1) primary, tested evidence based on published studies; 2) anecdotal, inferential evidence that could not be evaluated. This latter evidence has been separately described and is used merely to illustrate current trends in smoking attitudes and behaviours in the context of smoke-free legislation. It should not be taken as direct evidence on the effects of workplace smoking bans (see section 3.4 for further discussion of the way the evidence has been assessed). No Statement Grade Country/s Evidence WORKPLACE INTERVENTIONS Effectiveness of smoking cessation support and services to smokers in the context of a smoke-free workplace 1 Overall, it appears that workplace interventions in the context of environmental support (workplace smoking restrictions and educational campaigns) are effective in facilitating smoking cessation. One 2+ American study found that a smoking cessation programme delivered in the context of a workplace smoking ban and educational campaign produced long term success rates similar to smoking cessation programmes more broadly. However, another 2- American study found that environmental support may increase the success of workplace interventions, at least in the short term. One 1- study and one 2+study USA (Dawley et al. 1993 1-; Waranch et al. 1993 2+) Interventions in the workplace that work best 2 A 1++ systematic review and a 1+ meta-analysis of the available international literature indicates that the most effective smoking cessation interventions in workplace settings are those interventions that have proven effectiveness more broadly. There is strong evidence that group therapy, individual counselling and pharmacological treatments all have an effect in facilitating smoking cessation. However, both reviews failed to identify effects due to particular intervention type. There is also One 1++ systematic review and one 1+ metaanalysis International (Moher et al. 2005 1++; Fisher et al. 1990 1+) 5

evidence that minimal interventions including brief advice from a health professional are effective. Self help manuals appear to be less effective, although there is limited evidence that interventions tailored to the individual have some effect. 3 Two 1++ systematic reviews of international studies indicate that financial incentives are most commonly used by employers to encourage employee compliance with smoke-free workplace policies and the uptake of smoking cessation support. While the addition of incentives does not appear to increase the quit rates of smoking cessation interventions in the workplace, there is some evidence that such incentives do improve recruitment rates into worksite cessation programmes, which may lead to higher absolute numbers of successful quitters in the long-term. There is also some evidence that incentives may delay relapse to smoking, even if they don t prevent it altogether. 4 One 1++ systematic review of international studies found that adding a social or buddy support component to smoking cessation interventions in the workplace does not substantially improve cessation rates above and beyond group counselling and support alone. Two 1++ systematic reviews One 1++ systematic review International International (Moher et al. 2005 1++; Hey et al. 2004) (Moher et al. 2005 1++) 5 A 1++ systematic review and a 1+ meta-analysis of the available international literature suggests that intensive interventions are more effective than minimal interventions in facilitating smoking cessation in the workplace, although minimal interventions are more effective than no support at all. Extent to which the type of workplace and/or nature of employment influence smoking cessation outcomes 6 There are few available studies on how the type of workplace and/or nature of employment One 1++ systematic review and one 1+ metaanalysis One 2++ study International USA (Moher et al. 2005 1++; Fisher et al. 1990 1+) (Olsen et al. 1991 2++) Pp. 28 6

influence cessation outcomes in workplace interventions. One 2++ American study found that managers were more likely to successfully quit smoking in the long-term than smokers in other job categories. However, given that lighter smokers and older smokers were also more likely to quit successfully, the increased success of managers may have been due to baseline differences between themselves and smokers in other occupational categories (i.e. the managers are likely to have been older and lighter smokers). Variations in the effectiveness of workplace smoking interventions based on factors such as age, sex, gender, class or ethnicity 7 There is limited available evidence on how the effectiveness of workplace smoking cessation interventions varies with factors such as age, sex, gender, class or ethnicity. One 2++ American study that specifically explored the relationship between gender and smoking failed to find gender differences in quit rates following a smoking cessation intervention in the workplace. However, the study does note that significant gender differences were apparent in baseline smoking attitudes and behaviours. One 2++ study USA (Stockton et al. 2000 2++) Adverse or unintended outcomes of the intervention 8 A 1++ systematic review indicates that workplace interventions may have the potential for higher participation rates than other contexts, and also provide the opportunity to access smokers who would otherwise not be accessible. These represent significant potential outcomes of workplace interventions. 9 Evidence from a 2+ American study indicates that people who take part in workplace interventions in the context of smoking bans may enroll in the One 1++ systematic review One 2+ study International USA (Moher et al. 2005 1++) (Waranch et al. 1993 2+) 7

interventions to better control their cigarette consumption as opposed to intending to quit altogether. WORKPLACE POLICIES Influence of the smoke-free workplace policy on smoking patterns 10 A 1++ systematic review of international studies and two Irish studies (a 2++ study and a 2+ study) provide strong evidence that smoke-free workplace policies reduce the prevalence of smoking in the workplace and significantly reduce the exposure of employees to ETS at work. 11 The international evidence from a 1++ systematic review and a 1- meta-analysis on whether smoke-free work polices lead to a reduction in overall cigarette consumption is inconclusive. Reports from countries that have implemented national smoke-free legislation, including New Zealand (3+), the Republic of Ireland (3+), Italy (3+) and Norway (3-) indicate that a drop in cigarette sales has occurred although the true effect of smoke-free legislation on cigarette consumption is still to be determined. One 1++ systematic review, one 2++ study, one 2+ study One 1++ systematic review, one 1- metaanalysis, three 3+ reports and one 3- report International, Ireland International, New Zealand, Republic of Ireland, Italy, Norway (Moher et al. 2005 1++; Allwright et al. 2005 2++; Mulcahy et al. 2005 2+) (Moher et al. 2005 1++; Fichtenberg & S.A. Glantz 2002 1-; Allwright 2004 2+; Directorate for Health and Social Affairs 2005 3-; Thomson & Wilson 2006 3+; Gallus et al. 2006 3+) 12 The international evidence from a 1++ systematic review and a 1- meta-analysis on whether smoke-free work polices lead to a reduction in overall smoking prevalence is inconclusive. Monitoring data from Italy (rating 3+) indicates that a reduction in smoking prevalence appears to be occurring. However, given the recent implementation of national smoke-free legislation, analytic evidence on its effects on smoking prevalence is not presently available. One 1++ systematic review, one 1- metaanalysis, one 3+ report International, Italy (Moher et al., 2005 1++; Fichtenberg & S.A. Glantz, 2002 1-; Gallus et al. 2006 3+) 13 Although a 2++ American study and a 2+ Australian study have failed to find differences in the effectiveness of workplace smoking bans based on gender, One 1+ RCT, one 2++ study, and four 2+ studies USA, Australia, Finland (Moskowitz et al. 2000 2++; Owen & Borland 1997 2+; Heloma & 8

age, and ethnicity, three American studies (one 1+RCT and two 2+ studies) and a 2+ Finnish study indicate that bans may have a reduced impact on the smoking behaviours of people with a lower education and/or women of low socioeconomic status. Variations in the way that workplace policies are implemented that influence effectiveness 14 Evidence from a 2+ US study indicates that a total grounds ban may be more effective than an indoor smoking ban in reducing cigarette consumption. Impact of smoke-free workplaces on uptake of smoking cessation resources 15 According to a 2+ report from New Zealand, smoke-free legislation in conjunction with mass media campaigns does appear to lead to a statistically significant increase in phone Anecdotal evidence on smokefree legislation calls to telephone quitlines. Newspaper reports on the impact of smoke-free legislation in Scotland also indicate an increase in calls to the national telephone quitline in the period following the enactment of the legislation. 16 The available evidence from two 2+ US studies, one 3+ US report and one 2+ Australian study of workplace smoking bans indicates considerable variation in the impact of such bans on the demand for smoking cessation programmes. Demand for stop smoking programmes increased, although overall a relatively small proportion of smokers took advantage of the services that were provided. Anecdotal evidence on smokefree legislation Newspaper reports from Scotland indicate that demand for smoking cessation services has increased substantially since the enactment of the ban although it is unclear how much of this increase is due to the ban itself. One 2+ study One 2+ study Three 2+ studies and one 3+ case report USA New Zealand Scotland USA, Australia Jaakola 2003 2+; Farrelly et al. 1999 2+; Gritz et al. 1998 1+; Levy et al. 2006 2+) (Osinubi et al. 2004 2+) (Wilson et al., 2005 2+) (Howie et al. 2006) (Passannante et al. 1991 3+; Waranch et al. 1993 2+; Baile et al. 1991 2+; Borland 1990 2+) Scotland (Ross 2006; Brodie, 2006; BBC News, 2006) 9

17 A 2+ US study indicates that smoke-free legislation has a significant impact on NRT sales and substantially increases demand for these products. Anecdotal evidence on smokefree legislation 18 Anecdotal evidence Newspaper reports from Scotland also report an increase in demand for NRT following the enactment of the smoke-free legislation. Steps that should be taken prior to the introduction of smoke-free regulations to maximise the impact Workplace smoking bans and smoke-free legislation should be carefully planned, include the input of smokers, and be accompanied by provision of help and support for smokers. Public support for bans and legislation can be strengthened by using media campaigns to inform the public about the adverse health effects of passive smoking and by treating the issue as a worker protection law rather than a consumer protection law. An effort should be made to understand diversity and materials and messages should be culturally appropriate. An adequate revenue base is crucial to support the implementation of legislation. Acceptability of the policy to people affected by it 19 A 1++ systematic review on workplace interventions for smoking cessation finds consistent international evidence of positive behaviour and attitudinal changes following the implementation of workplace bans and restrictions. However, a 2+ Australian study and a 2+ UK survey find that based on indicators such as occupation and education, workplace bans are less acceptable to people of lower SES. One 2+ study One 1++ systematic review and two 2+ studies USA Scotland International, Australia, UK (Metzger et al., 2005 2+) (MacDonald, 2006; Rodrick, 2006) (Waranch et al. 1993; Anderson et al. 1999; Batlle et al. 1991; Baile et al. 1991; Moher et al. 2005; Strobl and Latter 1998); HDA 2005; R. Borland et al.. 2006a) (Moher et al., 2005 1++; Borland et al., 1990 2+; Lader & Goddard 2005 2+) 20 There is consistent international evidence from two 2+ studies and two 2- studies that there is less public support for a smoking ban in bars and pubs than in restaurants, although Two 2+ studies, two 2- studies, three 3+ studies Australia, USA, Italy, New Zealand, Republic of Ireland (Borland 2006b 2+; Friis & Safer, 2005 2-; Schofield & Edwards 1995 10

attitudes become more favourable following the implementation of such bans. Based on indicators such as education and occupation, people from lower SES groups appear to be the most opposed to bans in pubs/bars. However, the evidence on how attitudes to bans in restaurants and bars vary based on gender, age and ethnicity is inconclusive. 2+; Ahmed et al., 2004 2-) Monitoring data (3+) from countries that have implemented smoke-free legislation in bars and restaurants (New Zealand, Italy and the Republic of Ireland) indicates that support for legislation increases significantly following its implementation. 21 According to one 2+ Australian study and one 2- Swedish study, bar and restaurant managers (the former in particular) appear to have particularly negative attitudes towards smoking bans, substantially overestimating the potential financial impact of bans and underestimating the level of public support for them. However, according to the Australian study, attitudes become more favourable following the implementation of smoke-free legislation. One 2+ study, one 2- study, and one 3+ report Australia, Sweden, New Zealand (Jones et al.1999 2+; Hammar 2004 2-; Thomson & Wilson 2006 3+) Pp. 42-43 3+ monitoring data from New Zealand indicates that support for smoke-free bars amongst managers increases significantly following the introduction of legislation. Factors that affect compliance 22 According to monitoring data from the Republic of Ireland (3+), Italy (3+) and New Zealand (3+), compliance with smokefree legislation is reported to be extremely high. However, research from California (two 2++ studies) indicates that compliance in bars appears to be associated with their socioeconomic and Three 3+ reports, two 2++ studies Republic of Ireland, Italy, New Zealand, California (Office of Tobacco Control 2005 3+; Thomson & Wilson 2006 3+; Gallus et al. 2006 3+; Moore et al., 2006 2++; Lee et al. 2003 2++) 11

demographic characteristics. Adverse or unintended outcomes of the policy 23 According to four 2+ studies, workplace smoking bans and partial smoke-free legislation may lead to unhealthy smoking patterns such as compensatory smoking, harder smoking and displace smoking into the home. However, two 2+ studies indicate that compensatory smoking is unlikely to reach former levels and two 2+ studies indicate that smoking bans are associated with implementing a smoke-free home. 24 Overall, one 2- Australian study found that a workplace smoking ban was not a significant source of tensions between smokers and non-smokers, despite the minor advantages that were seen to be associated with exiled smoking. Six 2+ studies One 2- study Australia, USA, International Australia Pp. 43-44 (Wakefield et al. 1992 2+; Baile 1991 2+; Borland et al., 2006a 2+; Borland et al. 1999 2+; Adda & Cornaglia, 2006 2+; Chapman 1997 2+) Pp. 44-45 (Clarke et al., 1997 2-) Pp. 47 Inferential evidence The increased visibility of smoking that often accompanies the introduction of workplace smoking bans may lead to the stigmatisation of smokers and contribute to discriminatory practices and social stereotyping. (Greaves & Jategaonkar 2006) 25 Overall, one 2+ Scottish study and a 2+ study from the Republic of Ireland indicate that smoke-free legislation may encourage smokers to congregate around building entrances and exits, thereby increasing the exposure of nonsmokers to second-hand smoke through more intensive contact with the smoke as they enter buildings and drifting smoke issues. One 2+ study, one 2+ study Scotland and Republic of Ireland (Parry et al. 2000 2+; Mulcahy et al. 2006 2+) Pp. 47-48 26 Two 2- English studies and one 2+ Scottish study report that workplace smoking bans may lead to an increase in dangerous smoking practices (such as smoking in inappropriate locations and the build-up of smoking related debris). One of the English studies also raises the Two 2- studies and one 2+ study UK (Strobl & Latter 1998 2-; Parry et al. 2000 2+; Anderson 1991 2-) Pp. 49 12

potentially negative effects of bans on smokers who must venture outside to smoke, even in poor weather conditions. 13

2. Background Cigarette smoking is the leading cause of preventable death in the United Kingdom today. In England alone, between 1998-2002 smoking was estimated to be responsible for 86,500 deaths per year (Twigg et al. 2004). More than half of all smoking related deaths were due to respiratory diseases such as lung cancer, chronic obstructive pulmonary disease (COPD) and pneumonia, while ischaemic heart disease, other cancers, circulatory and digestive diseases accounted for the rest (Royal College of Physicians 2000). However, although the harms caused by cigarette smoking are well established, there is a growing body of evidence that environmental tobacco smoke (ETS), otherwise known as second-hand smoke or passively ingested smoke, also causes harm to those exposed to it. The first study linking passive smoking and lung cancer was published in 1981 (Hirayama 1981) and since that time there has been a groundswell of literature on the health-related harms connected with ETS. In a recent assessment of the available evidence, the Scientific Committee on Tobacco and Health (SCOTH 2004) concluded that exposure to ETS substantially increases the risk of lung cancer and ischemic heart disease amongst non-smokers, and children exposed to ETS are at increased risk of bronchitis, asthma attacks, pneumonia, a reduction in lung function, middle ear disease and Sudden Infant Death Syndrome (SIDS). Given that approximately one quarter of Britons smoke (Lader and Goddard 2005), exposure to ETS remains a significant issue. Indeed, a recent study (Jamrozik 2005) estimates that across the United Kingdom as a whole, passive smoking in the workplace is likely to be responsible for the death for more than two employed people per working day (617 deaths per year), including 54 deaths in the hospitality industry each year almost three times the number of deaths from industrial injuries and accidents (Health and Safety Commission 2003). Although workplace restrictions on smoking have been contentious and entail arguments about freedom, privacy, workplace health and rights, as the health impacts of ETS have become more clearly established, the movement to ban smoking in workplaces has gained momentum worldwide, with a range of countries, states and cities implementing smoke-free legislation. In the United States, Canada and Australia, a growing number of regions have established smoke-free ordinances. Two of the most extensive and well publicised examples of such ordinances can be found in California, which extended its smoke-free legislation in January 1998 to include bars, lounges and nightclubs, and New York, which implemented a Smoke Free Air Act in 2003 prohibiting smoking in virtually all enclosed public places. More recently, several countries have implemented comprehensive smoke-free legislation at the national level, including the Republic of Ireland in March 2004, Norway in June 2004, New Zealand in December 2004, Italy in January 2005, Sweden in June 2005, and Scotland in March 2006. The decision of its closest neighbours to go smoke-free has placed the English government under considerable political pressure to implement similar legislation. 2.1 Current Smoking Restrictions in England In 1998, following the publication of the landmark White Paper Smoking Kills (Department of Health 1999), the health and economic effects of smoking were placed upon the political agenda in England as never before. Smoking Kills laid out a 14

comprehensive plan for reducing the prevalence of smoking in the UK, and entailed measures such as a ban on tobacco advertising, increases in the price of tobacco, a significant injection of funding into smoking cessation services and strategies to reduce smoking in work and public places (McNeill et al. 2005). Specifically, the White Paper indicated that the Health and Safety Commission (HSC) would consult on a new Approved Code of Practice on smoking in the workplace in order to substantially toughen existing measures by defining the kind of smoking policies employers would need to implement to comply with existing health & safety legislation. It also included provision of a Public Places Charter whereby signatories representing key organisations in the hospitality industry agreed to increasing [the] provision of facilities for non-smokers and availability of clean air. In March 2000 formal targets were agreed between industry representatives and the Department of Health with compliance to be achieved by January 2003 (The Charter Group 2003): 50% of all pubs and half the members of the Restaurant Association should have a formal smoking policy and carry an external policy sign 35% of these premises should restrict smoking to designated and enforced areas and/or have ventilation that met the agreed standard. Although The Public Places Charter on Smoking Industry Progress Report (The Charter Group 2003) found that both of the targeted sectors (pubs and restaurant association members) had comfortably exceeded their targets 1, it also concluded that non-smoking pubs remain a rarity (The Charter Group 2003). Moreover, while just over half of the respondents surveyed in the 2004 Smoking-related Behaviour and Attitudes Survey (Lader and Goddard 2005), indicated that they worked in locations where smoking was not allowed at all on the premises, a further 37% worked in premises that had smoking in designated areas and 8% worked in locations where there were no restrictions on smoking at all (see table 1). Table 1. Restrictions on smoking where respondent currently works, 1996-2004 Level of restriction 1996 1997 1999 2000 2001 2002 2003 2004 % % % % % % % % No smoking at all 40 42 48 44 47 50 50 51 Designated areas only 42 41 37 40 38 36 38 37 No restrictions at all 13 13 11 11 9 9 8 8 Don t work with others 5 4 4 5 6 5 4 4 Base=100% 2154 2195 2104 1883 2040 2251 2084 2174 Although the Smoking-related Behaviour and Attitudes Survey (Lader and Goddard 2005) shows that there has been a substantial increase in the implementation of workplace smoking bans since 1996 (see table 1), it is clear that voluntary legislation has not led to the uniform or universal creation of smoking restriction policies, particularly in pubs and restaurants. Moreover, the Smoking-related Behaviour and Attitudes Survey (Lader and Goddard 2005) has also revealed that men, heavy smokers and those in routine and manual occupations are the most likely to work at premises that do not restrict smoking. 1 Questions can be raised about the utility of these targets in increasing the availability of clean air in pubs and restaurants as they merely required half of these establishments to have a formal policy in place, and a third of these locations to restrict smoking to designated areas. 15

2.2 Smoking-Related Inequalities In England, as in most developed countries, inequalities in mortality and morbidity are strongly linked to socio-economic factors such as social class (occupation), income, level of education and area of residence (Chesterman et al. 2005). Socioeconomic inequalities in smoking are one of the main factors responsible for the gap in life expectancy between rich and poor in England (Killoran et al. 2006). Indeed, tobacco is responsible for more than half the difference in male mortality between those in the highest and lowest socio-economic groups living in the UK (Jarvis and Wardle 1999). Moreover, the gap between the smoking rates of manual and non-manual groups in the UK appears to be widening as there has been a sharper decline in smoking prevalence among non-manual compared with manual social groups (Killoran et al. 2006). The fact that people in routine and manual occupations are less likely to work in environments that restrict smoking means that whether they smoke or not, their exposure to ETS at work is higher than for those people in intermediate and professional and managerial occupations. Furthermore, this increased exposure to ETS also seems to carry over into the leisure environments of people from routine and manual groups. A recent study (Tocque et al. 2005) has found that the pubs and restaurants most likely to allow unrestricted smoking are situated in the areas of highest deprivation. These findings provide further evidence that the voluntary legislation has not protected all workers equally from the effects of ETS and workers in some occupations (particularly routine and manual occupations) continued to suffer high levels of exposure. 2.3 Smoke-free Legislation in England In response to these factors and ongoing lobbying by health campaigners, in 2004 the Choosing Health White Paper (Department of Health 2004) announced the Government s proposed action to introduce smoke-free workplaces through a stepped approach: 1) by the end of 2006, government departments and the NHS will be smoke-free 2) by the end of 2007, all enclosed work and public places, other than licensed premises 3) by the end of 2008, arrangements for licensed premises in place The proposed legislation was brought forward in the Health Bill in 2005; however, as a result of a strong view amongst members of both the public and parliament that this legislation did not go far enough, the Government brought forward alternative options for extending the smoke-free provisions, including the possibility of national legislation to make all indoor public places and workplaces completely smoke-free (with minimal exemptions) (Department of Health 2006). This option was resoundingly favoured by the House of Commons and the legislation is due to be implemented in 2007 (Department of Health 2006). The successful campaign for comprehensive smoke-free legislation in England represents a significant achievement for the tobacco control movement, and a turning point in the development of a national, comprehensive tobacco control policy and has been described as the single most important public health measure of the past 30 years (Willmore 2006). 16

Through this legislation, the government s objective is to: reduce the risks to health from exposure to second-hand smoke recognise the right to be protected from harm and to enjoy smoke-free air increase the benefits of smoke-free enclosed public places and workplaces for people trying to give up smoking so that they can succeed in an environment where social pressures to smoke are reduced save thousands of lives over the next decade by reducing both exposure to hazardous second-hand smoke and overall smoking rates (Department of Health 2006). Although the ostensible goal of smoke-free legislation is to reduce the health effects of ETS on workers, a reduction in smoking prevalence is expected to be an ancillary benefit. However, given that smoke-free legislation operates simultaneously within a comprehensive tobacco policy context, it is extremely difficult to disentangle the specific impacts of each policy on populations or sub-populations (Greaves et al. 2004). As G.T. Fong et al. (2006) point out, Evaluation of tobacco control policies at the population level is in its early stages of development, and as such, studies on the effectiveness of tobacco control policies suffer from design limitations. Cross-sectional studies are generally weak in their ability to yield causal attributions. Longitudinal studies are higher in internal validity, but the limited number of such studies in tobacco policy research often lack comparison groups, and are thus unable to disentangle policy effects from secular trends and historical event threats to internal validity. Thus, while policy makers are understandably hopeful that bans implemented as part of a comprehensive tobacco control policy will have a clear effect on cigarette consumption, prevalence rates and health inequalities, unthreading the effects of smoke-free legislation is virtually impossible. Ultimately, while it is clear that comprehensive tobacco control policies have improved overall population health and health inequalities, there is a lack of research into the various components of tobacco policies and their effectiveness (Greaves et al. 2004). This is particularly true of smoke-free legislation; given its very recent implementation in countries around the world, its long-term impact is still to be assessed. Further research is also needed on the impact of such legislation on subpopulations, particularly disadvantaged groups. As L. Greaves et al. (2004) point out, the actual patterns, trends and effects of policy implementation among disadvantaged groups are often undiscovered, preventing due consideration of the effects of interventions, and thus, policies, on many sub-populations. A key aim of this review is to provide an initial analysis of the existing evidence on the effects of policy implementation with a particular focus on deprived populations. It contains assessments of the available data on the extent to which workplace policies stimulate, support and utilise smoking cessation, with a view to determining the likely effects of the introduction of national smoke-free legislation in England in 2007. 17

3. Methodology 3.1 Literature Search Julie Glanville and Kate Light (Centre for Reviews and Dissemination, University of York) conducted the searches for this rapid review in May 2006, with input from NICE and the British Columbia Centre of Excellence for Women s Health (BCCEWH) team. The first literature search covered systematic reviews in the standard databases and produced 533 records (see Appendix C, parts 1a & 1b). The second literature search covered non-reviews in the standard databases and produced 6878 records (see Appendix C, part 2). A further Medline search of both reviews and non-reviews was undertaken using the earlier Medline search strategy, but changing line 18 to include abstracts as well as titles (see Appendix C, part 3). This search produced 740 records (reviews) and 4872 records (non-reviews) respectively. In total the BCCEWH team received 13,023 references. A detailed report of processes, databases, and search terms used in the review is presented in Appendix C. Studies not published in English were excluded from the review. 3.2 Selection of Studies for Inclusion Once the literature search was complete the project team selected relevant studies based on the criteria outlined in section 4.1 of the Public Health Guidance Methods Manual. Before acquiring papers for assessment, preliminary screening of the literature search was carried out to discard irrelevant material. Titles were initially scanned by one reviewer who removed the clearly irrelevant studies. The remaining 200 abstracts were independently scrutinised in relation to the research questions by two reviewers and those that did not directly deal with the issues raised in the research questions were eliminated. Once this sifting process was complete, paper copies of the selected studies or reviews were acquired for assessment. 3.2.1 Policies and Interventions The review was international in scope and included workplace smoking policies and workplace smoking cessation interventions aimed at the workforce. Policies of interest ranged from total indoor bans to total grounds bans on smoking in the workplace. Interventions of interest were those provided in the workplace. All types of intervention were considered, such as group therapy, individual counselling, selfhelp materials, and nicotine replacement therapy (NRT). 3.3 Outcomes The outcomes of interest were: 1. Changes in attitudes toward the policy or intervention from the employers and/or employees point of view 2. Changes in the consumption of cigarettes at the workplace following the policy or intervention (with biochemical validation where recorded). The research team was particularly interested in literature that analysed these outcomes based on setting and factors such as sex, gender, class, ethnicity and age. A recent Cochrane Review (Moher et al. 2005) on the effectiveness of workplace smoking interventions in reducing cigarette consumption provided a key source of information in the following review. When evaluating the effects of workplace smoking cessation interventions and bans on smoking attitudes and behaviour, the Cochrane Review has been used as the key source of evidence rather than attempting to summarise all of the individual studies identified in the literature search 18

on this topic. 2 However, a number of the studies listed in the Cochrane Review have been separately considered in the following report. The procedure used to determine whether studies listed in the Cochrane Review should be independently examined was as follows: 1) The abstracts of all the studies listed in the review were searched. 2) Any abstracts that provided mention of the effects of policies or bans on any of the socio-demographic variables of interest (such as gender, class, etc) or any of the other issues of interest (such as unintended consequences, etc) were noted and the copies of the full studies obtained. 3) Following scrutiny of the full papers, two reviewers independently determined whether the studies were relevant enough to rate as evidence. Following the elimination of 12,823 irrelevant records based on title alone, the two reviewers assessed abstracts of 200 records for possible inclusion and 57 records were determined to be addressing the key outcomes and populations of interest based on their abstract. Full copies of these studies were obtained and were independently assessed for inclusion by two reviewers. Of these studies, 44 met the inclusion criteria for this rapid review (see figure 1). A list of excluded studies with reasons for exclusion is presented in Appendix A. Figure 1. The evidence 13,023 studies identified in literature search 12,823 irrelevant sources 200 relevant sources 57 sources assessed 143 sources for inclusion scouted as background 44 studies met inclusion criteria 17 studies excluded from review 1 study used as background 3.4 Quality Appraisal Tested Evidence All of the studies that met the inclusion criteria were rated by two independent reviewers in order to determine the strength of the evidence. Once the research design of each study was determined (using the NICE algorithm), studies were assessed for their methodological rigour and quality based on the critical appraisal checklists provided in Appendix B of the Public Health Guidance Methods Manual (see table 1). Each study was categorised by study type and graded using a code 2 Obviously there was a substantial degree of overlap between those studies identified in the literature search and those listed in the Cochrane Review. 19

++, + or, based on the extent to which the potential sources of bias had been minimised. Those studies that received discrepant ratings from the two reviewers were given to a third reviewer for final evaluation. Several qualitative studies were included in the review; while the Public Health Guidance Methods Manual provides guidance on how their methodological rigour should be assessed, the research team had to make a judgement about the level the qualitative studies should be assigned, based on their relevance to the research question. The research team decided that qualitative studies containing rigorous analysis of cigarette consumption and smoking prevalence should receive a level 2 rating. There is currently no methodological checklist for cross-sectional studies in the Public Health Guidance Methods Manual. In order to assess the quality of these studies, modifications to existing NICE checklists are recommended and a crosssectional checklist based on the cohort study checklist in the manual was created (see Appendix D). Table 2. Level and quality of evidence Type and quality of evidence 1 ++ High quality meta-analyses, systematic reviews of RCTs, or RCTs (including cluster RCTs) with a very low risk of bias 1 + Well conducted meta-analyses, systematic reviews of RCTs, or RCTs (including cluster RCTs) with a low risk of bias 1 Meta-analyses, systematic reviews of RCTs, or RCTs (including cluster RCTs) with a high risk of bias 2 ++ High quality systematic reviews of these types of studies, or individual, non- RCTs, case-control studies, cohort studies, CBA studies, ITS, and correlation studies with a very low risk of confounding, bias or chance and a high probability that the relationship is causal 2 + Well conducted non-rcts, case-control studies, cohort studies, CBA studies, ITS and correlation studies with a low risk of confounding, bias or chance and a moderate probability that the relationship is causal 2 Non-RCTs, case-control studies, cohort studies, CBA studies, ITS and correlation studies with a high risk or chance of confounding bias, and a significant risk that the relationship is not causal 3 Non-analytic studies (for example, case reports, case series) 4 Expert opinion, formal consensus Grading the evidence ++ All or most of the quality criteria have been fulfilled Where they have been fulfilled the conclusions of the study or review are thought very unlikely to alter + Some of the criteria have been fulfilled Where they have been fulfilled the conclusions of the study or review are thought unlikely to alter - Few or no criteria fulfilled The conclusions of the study are thought likely or very likely to alter 20

Applicability of evidence No statements have been made about the applicability of the evidence to the UK setting. The BCCEWH team did not feel that it was appropriate to discuss the applicability of the evidence for two reasons. First, the review was international in scope and covers studies largely from Australia, the United States and Canada. It is therefore unclear how readily the findings of these studies translate to an English setting. Second, as much of the tested evidence relates to localised workplace smoking bans and jurisdictions with partial smoke-free legislation in place, it is also unclear how applicable these findings are when attempting to determine the likely effects of large-scale clean air laws and smoke-free legislation such as the legislation to be implemented in England in 2007. Strength of evidence Although tested evidence from published studies provides the primary source of evidence for this review and statements regarding the strength of the evidence have been drawn solely from this material, given that it is unclear how readily this evidence translates to countries that have implemented national smoke-free legislation, an effort has been made to consider the available level 3 evidence of the effects of smoke-free legislation from countries that have recently gone smoke-free (e.g. the Republic of Ireland, Norway, Italy and New Zealand). Although level 3 evidence has not been listed in the evidence table because it could not be evaluated accorded to established standards of evidence, some attempt has been made to rate the general quality of this evidence (see Appendix B for a full list of level 3 sources and rationale for their quality rating). These level 3 reports do not constitute analytic evidence as they merely summarise monitoring data on the legislation, without attempting to account for secular trends that may have influenced present smoking attitudes and behaviour. However, while this data should not be taken as definitive evidence, it does provide the only available information on the effects of national smoke-free legislation and has therefore been outlined in the review alongside of the tested evidence. Other evidence Aside from the tested and inferential (level 3) evidence outlined above, this review also draws on anecdotal evidence emerging from Scotland based on its recent experience with going smoke-free as well as anecdotal reports about the steps that can be taken prior to introducing legislation to maximise its effects. This anecdotal evidence is should not be taken as definitive evidence on the effects of smoke-free legislation; it has been separately sign-posted in the evidence statements as anecdotal evidence for this reason). Finally, one section of the report also discusses inferential evidence (see section 4.2.7) where the research team has inferred possible effects of legislation, rather than drawing on tested evidence. 3.5 Synthesis Due to heterogeneity of design among the studies, a narrative synthesis was conducted. 21

4. Summary of Findings This review focuses on workplace policies to stimulate, support and utilise smoking cessation in England. A workplace policy is defined as a complete prohibition on smoking, with or without clinical support for cessation attempts. A workplace is defined as all enclosed or substantially enclosed settings where people are employed, including areas they have to enter as part of their paid or unpaid employment 3 (e.g. office buildings, factories, hospitals, leisure facilities, bars and restaurants). The research questions devised for this review relate to two distinct categories: workplace interventions for smoking cessation and workplace smoking restrictions. Questions about Interventions 1) What is the effectiveness of smoking cessation support and services to smokers in the context of a smoke-free workplace? 2) Which interventions in the workplace work best? What is the content of intervention? Does the effectiveness depend on the job title/position of the deliverer? Does the intensity of the workplace intervention influence smoking cessation outcomes? 3) To what extent does the type of workplace and/or nature of employment influence smoking cessation outcomes? 4) How does the effectiveness of workplace interventions vary with factors such as age, sex, gender, class or ethnicity? 5) What are the adverse or unintended outcomes of the intervention? 6) How acceptable is the intervention to those affected by it? 3 We have excluded schools from this review, because none of the available studies on school smoking policies focus on the effects of these policies on staff and visitors. Rather, they aim to reduce the uptake of smoking among adolescents, which is outside the scope of this review. 22

Questions about Policy 1) What is the effectiveness of the smoke-free workplace policy in: encouraging quit attempts decreasing tobacco consumption decreasing overall smoking prevalence decreasing the prevalence of smoking in the workplace 2) How does the way that workplace policies are implemented influence effectiveness? 3) What is the impact of smoke-free workplaces on uptake of smoking cessation resources? For example, NHS Stop Smoking Services, telephone quitlines, etc. 4) What steps could be taken prior to the introduction of smoke-free regulations to maximise the impact? E.g., public information campaigns, increased resources for smoking cessation support. 5) How acceptable is the policy to people affected by it? Based on the setting of the policy (workplace, restaurants, bars) Based on factors such as age, sex, gender, class or ethnicity 6) What factors affect compliance? 7) What are the adverse or unintended outcomes of the policy? 23